This is a simple demo of an API written in Node.js, hosted in Azure Functions, and uses DocumentDB as a database. In other words, there is no server, no database to set up. Just deploy and run.
This is a basic implementation of a server that tracks the current version of systems (nodes) deployed in various locations. Clients regularly check in, and they're told if there is a newer version of their software available.
This is only a sample of some of the concepts. Not a complete solution by any means.
Why maintain a server if you don't have to? Azure Functions lets us focus 100% on writing our code. Azure functions worries about deployment, execution, and scaling. Every day it gets more powerful and more sophisticated.
Document databases let us store JSON objects (documents) in a direct, natural way. No need to map our fields to a database and jump through unnecessary steps.
Okay, so why DocumentDB specifically? Simple, it was born in the cloud, designed to scale infinitely, and most importantly it's 100% hosted and managed for us. It's an excelent complement to Azure Functions.
Initially, I used the DocumentDB native client, but it was a little more verbose than I liked. Fortunately, DocumentDB also supplies a MongoDB compliant interface, and then you can use the standard MongoDB client. For this sample, it was much more simple to use.
Running an Azure function locally is as easy as running func run .\ListAllNodes\ where ListAllNodes is the name of your function.
It took me 60 seconds to deploy this to Azure. Simply create an Azure Function App in the portal. In my case I picked the "Consumption Plan" hosting plan option, because you're essentially paying for each request.
After creating the function app, go to the app settings.
In the Deploy section, click "Configure Continuous Integration".
I configured mine to deploy from this GitHub repo. (instructions omitted)
The last set was to configure the App Settings screen and set up the required application settings. Namely, the dbUrl parameter.
